Relaxed Equivariant Graph Neural Networks
This addresses a specific challenge in physics and chemistry modeling for researchers using equivariant neural networks, but it appears incremental as it builds on the existing e3nn framework.
The paper tackles the problem of modeling symmetry breaking within continuous groups in 3D Euclidean equivariant neural networks by introducing a framework with relaxed weights, showing empirically that these weights learn the correct amount of symmetry breaking.
3D Euclidean symmetry equivariant neural networks have demonstrated notable success in modeling complex physical systems. We introduce a framework for relaxed $E(3)$ graph equivariant neural networks that can learn and represent symmetry breaking within continuous groups. Building on the existing e3nn framework, we propose the use of relaxed weights to allow for controlled symmetry breaking. We show empirically that these relaxed weights learn the correct amount of symmetry breaking.